Thyroid cancer around Fukushima

[overtone]

One of the reasons the signal from Chernobyl hid for so many years was that the data was not interpreted as checking a hypothesis but as detecting a signal in the absence of the hypothesis.

 

If one samples a large population for a disease and finds a cluster of cases, the probability of it being chance clustering is one number. If one specifies in advance a particular group of people from within this population to exhibit that particular clustering for some reason, and they do, the probability of it being chance clustering is a much different number.